Curated interview practice questions with answers. These are practice scenarios, not a verified list of past interview questions.
About NVIDIA hiring for Software Testing roles
NVIDIA designs GPUs and accelerated computing platforms used in gaming, data centres, AI and automotive. It also develops software like CUDA and AI frameworks. Its India offices in Pune, Bengaluru and Hyderabad work on chip design, verification, system software and AI. It hires for software engineers, verification engineers and test engineers.
For Software Testing roles, this NVIDIA practice set focuses most on Agile, SQL and API Testing (10), Bug Life Cycle and Jira (6), Automation Basics for Manual Testers (6), Testing Types (5). It has 3 easy, 32 medium, 5 hard questions, so you can start easy and work up to the harder ones.
Company type
Hardware & semiconductors
Headquarters
Santa Clara, California, USA
Main India offices
Pune, Bengaluru, Hyderabad
NVIDIA Software Testing interview process
Online assessment or technical screening
Technical round 1
Technical round 2
Technical round 3 or manager round
HR round
Typical pattern based on public candidate experiences. Rounds vary by role, team and year.
Solved for NVIDIAStart practicing
3 QUESTIONS FREESign up to unlock the full question bank and track your progress.Unlock every question and answer, then continue practising company-wise and package-wise.
Defect age is how long a defect stays open, measured either in time, from the date it was logged to the date it was closed, or in phases, from the phase where it was introduced to the phase where it was found. High defect age signals slow fixing or triage. Phase-based age shows how late defects are caught; a requirement defect found in UAT has a high age and is costly.
Re-check environment, data, browser and steps; check the logs; try the reporter's setup; add video and console logs; mark it 'cannot reproduce' only after discussing with the developer, and keep monitoring.
05
Your automated tests pass one by one but fail randomly when run in parallel. How do you fix the test data so tests are independent?
Each test should create the data it needs, usually through an API or database helper, use unique values such as an email with a timestamp, and clean up afterwards. Avoid tests that depend on data made by another test or on a shared account other tests change. For fixed reference data, seed it before the suite runs. Independent data allows tests to run in parallel and in any order without random failures.
06
What is the use of SQL GROUP BY and HAVING in test validation?
Model answer: Situation: our sprints kept ending with stories stuck in testing. Task: suggest a fix in the retrospective. Action: I showed that most stories reached QA in the last two days and proposed smaller stories, early test case sharing and developers deploying to QA daily. Result: the team tried it next sprint, and the number of stories carried over dropped from five to one.
A test oracle is the source you use to decide whether a test result is correct, that is, the expected result. It can be a requirement document, a user story, an older version of the system, a calculation in a spreadsheet, a domain expert or common sense. For example, when testing a tax calculator, the official tax slab table is the oracle. Without a reliable oracle, you cannot say a test passed or failed.
A deep link opens a specific screen inside the app from a URL, email, SMS, QR code or notification. Test that each link opens the correct screen with the right data when the app is installed, closed or running in background. If the app is not installed, it should open the store page or a web fallback. Test links that need login, expired or invalid links, and that parameters cannot be tampered with to access other users' data.
12
An automated test passes on your laptop but fails in the CI pipeline. How do you troubleshoot?
Check the CI logs, screenshots, video or trace to see exactly where it failed. Then compare the environments: headless versus headed browser, screen size, browser version, time zone, environment URLs and secrets, test data and network speed. Run the test locally in headless mode with the CI settings. Common causes are timing, missing data in the CI environment and tests depending on run order. Fix the cause, not with a sleep.
13
What is the difference between a defect and a production incident?
A defect is a flaw in the software that makes it behave differently from the requirement, whether found in testing or production. A production incident is an event that disrupts or degrades service for users, such as an outage or a failed payment flow. An incident may be caused by a defect, but also by infrastructure, configuration or third-party failures. Incidents are handled by an incident process first, then linked to defects for the fix.
14
SQL: verify that soft-deleted users do not appear in the active users list.
-- the active list query used by the app should return 0 rows here
SELECT u.user_id, u.email
FROM users u
WHERE u.is_deleted = 1
AND u.user_id IN (
SELECT user_id FROM active_users_view
);
Sprint planning (estimate test effort), daily stand-up (status and blockers), sprint review (demo) and retrospective (improve). Testers clarify acceptance criteria early and test within the sprint.
Valid keyword, partial match, no results, special characters, very long text, empty search, case sensitivity, spelling errors, suggestions and speed of results.
17
How do Selenium, Playwright and Cypress differ at a high level?
Selenium is the oldest, uses the WebDriver standard, supports many languages like Java, Python and C# and every major browser, and is common in older enterprise suites. Playwright supports TypeScript, Python, Java and .NET, runs Chromium, Firefox and WebKit, and has built-in auto-waiting, tracing and parallel runs. Cypress runs inside the browser with JavaScript or TypeScript, has a friendly runner and is popular with front-end teams.
18
What are transaction isolation levels and what is a dirty read?
Isolation levels control how much one transaction can see of another's uncommitted work. Common levels are Read Uncommitted, Read Committed, Repeatable Read and Serializable. A dirty read is reading data another transaction changed but has not committed; if it rolls back, you acted on data that never existed. Testers check isolation using two database sessions: update without committing in one and query in the other to see what is visible.
19
The UI shows the correct order total, but the database has a different value. What do you check?
Note the order ID and compare the UI value, the API response and the database value to find where they differ. Check if the UI calculates the total on the screen instead of using the saved value, whether taxes, discounts or rounding are applied at different layers and whether a later update job changed the record. Log the bug with all three values, because invoices and reports will use the database figure.
20
Write test cases for live order tracking in a food delivery app.
Verify each status change appears in order: placed, accepted, preparing, picked up and delivered. Check the rider's location updates on the map at the expected interval. Test the estimated time updates when there is a delay. Verify notifications for each status. Check behaviour when the network drops and recovers, when the app is reopened, and when the order is cancelled midway. Confirm tracking stops after delivery and the rating screen appears.
21
Explain three usability heuristics with examples from a web app.
Visibility of system status: after clicking Pay, show a progress indicator so users do not click again. Error prevention: disable the Submit button until required fields are valid, or ask for confirmation before deleting an account. Recognition rather than recall: show recently searched items and saved addresses instead of making users remember them. Other heuristics include consistency, user control with undo, and clear help for recovering from errors.
Test with location permission allowed, denied and approximate only. Use mock location tools to simulate different cities and moving routes. Verify the app shows nearby results correctly, updates location while moving and handles GPS turned off. Check accuracy indoors and with weak GPS, battery impact of continuous tracking and that background location is used only if needed and explained to the user.
Check that every field has a visible label tied to it, so clicking the label focuses the field and a screen reader announces it. Required fields must be indicated in text, not only by colour or an asterisk without explanation. Submit with errors and check that messages are specific, appear next to the field, are announced by the screen reader and focus moves to the first error. Also check autocomplete for name, email and address.
28
How do you test biometric login with fingerprint or face unlock?
Test enabling biometrics in app settings after a password login. Verify login with a registered fingerprint or face, and failure with an unregistered one, falling back to PIN or password after the allowed attempts. Remove or add a fingerprint in device settings and check the app asks for re-authentication. Test devices without biometric hardware, disabled biometrics and app behaviour after an OS update or reinstall.
Status code, response body and schema, headers, response time, error messages for invalid input, authentication (401/403) and that data is saved in the database.
30
Which basic Git commands should a manual tester learn before working on automation?
Start with: git clone to copy the test repository; git pull to get the latest changes; git checkout -b feature/login-tests to create a branch; git status and git diff to see changes; git add and git commit -m to save work with a message; and git push to send the branch so you can raise a pull request for review. Also learn to resolve simple merge conflicts and never commit passwords.
31
Tell me about a time exploratory testing helped you find a critical bug.
Model answer: Situation: a new leave management module had passed all scripted cases. Task: I had a one-hour exploratory session. Action: I applied leave, then changed my manager in the profile while the request was pending. Result: the request disappeared for both managers and could never be approved. It was fixed before release, and the team added role-change scenarios to the regression suite.
Backward compatibility testing checks that a new version still works with older versions of things it depends on or that depend on it. Examples: a new app version opens files and data created by the old one, a new API version still serves older mobile app versions that users have not updated, and an upgraded database keeps old records readable. It prevents breaking existing users when only some parts of the system are upgraded.
33
The developer says "it's not a bug". What do you do?
Pull request checks must be fast, usually under ten to fifteen minutes, and highly reliable, so they include unit tests, API tests for changed services and a small UI smoke set of critical flows. The full UI regression, cross-browser runs, long data-driven suites and performance checks run nightly or before release. Tag tests by suite, measure run times and flakiness, and move tests between suites based on value and speed.
35
What is localisation and internationalisation testing?
Internationalisation tests that the product can support many languages and formats; localisation tests that a specific locale is correct for text, currency, date and layout.
Send requests with missing required fields, wrong data types, values beyond limits, invalid formats such as a bad email, unknown fields and an empty or malformed JSON body. The API should return a 4xx code such as 400 or 422, a clear error message that names the field, and no stack trace or database details. Also check that nothing was partly saved in the database after the failed request.
Functional testing checks what the system does (login, checkout). Non-functional testing checks how well it does it (performance, security, usability, compatibility).
39
SQL: verify that row counts and totals match between a source table and a target table after a data load.
-- Compare counts and sums side by side; both columns must match
SELECT
(SELECT COUNT(*) FROM staging_orders) AS src_count,
(SELECT COUNT(*) FROM dw_orders) AS tgt_count,
(SELECT SUM(amount) FROM staging_orders) AS src_amount,
(SELECT SUM(amount) FROM dw_orders) AS tgt_amount;
Set browser zoom to 200 percent and check that all text is readable and no content or buttons are cut off. Then zoom to 400 percent, or use a 1280 pixel wide window, which gives a 320 CSS pixel viewport; content should reflow into one column without horizontal scrolling, except for things like data tables and maps. Also increase text spacing and the operating system font size to check that labels do not overlap.
Tips to crack the NVIDIA Software Testing interview
Revise C and C++, operating systems and computer architecture, which are core to most NVIDIA engineering roles.
Learn GPU and CUDA basics, including parallel programming, threads, blocks and the GPU memory hierarchy.
Read about NVIDIA's data centre and AI platforms to understand where much of its current work is focused.
Keep a one-minute story ready for every project on your CV: problem, your part, tools, result.
Practise the hard questions on this page aloud; NVIDIA interviewers usually ask follow-ups on your first answer.
Frequently asked questions
How many rounds are there in the NVIDIA Software Testing interview?
Candidates usually go through 5 stages: Online assessment or technical screening; Technical round 1; Technical round 2; Technical round 3 or manager round; HR round. The exact rounds change with the role, team, location and your experience, so confirm the process with the recruiter.
Which topics should I prepare for the NVIDIA Software Testing interview?
This practice set leans on Agile, SQL and API Testing (10), Bug Life Cycle and Jira (6), Automation Basics for Manual Testers (6), Testing Types (5). Revise those first, then practise explaining one or two of your own projects end to end, because most NVIDIA interviewers ask you to walk through real work.
Are these the exact questions asked at NVIDIA?
No. These are curated practice questions that match the skills and difficulty usually tested in hardware & semiconductors interviews like NVIDIA's. They are not a leaked or verified list of past questions.
Where does NVIDIA hire Software Testing professionals in India?
NVIDIA has large teams in Pune, Bengaluru, Hyderabad. Openings and locations change often, so check the NVIDIA careers page and filter by city and experience.